Moore's Law, Cloud Computing and Business Intelligence

DW/BI and Cloud Computing

As we all know, Data Warehousing and Business Intelligence has been the key to making informed strategic decisions for an organization’s success. DW/BI are generally hosted on servers within the organization so as be accessible to a network of decision making executives.  In order to meet the increasing demand in data analysis, the varied functionalities expected from the DW/BI tools, and the enormous data the BI tools handle, DW/BI has started to move from servers and local machines to the cloud.

Cloud Computing is big advancement in technology is the cloud. Cloud provides the biggest advantage of flexibility to organizations to meet with their fluctuating demands. Organizations are moving to cloud in order to meet with their requirements in scalability, computing resources, speed, balancing workloads, storage and reliability. In addition to providing these benefits, it provides an effective cost-saving option as it meets the need of the house with flexibility in resources being used depending on the requirement.

With this introduction on DW/BI on cloud, let’s get into some of the latest advancements that combines DW/BI with cloud computing.

1.       Hautelook with Amazon RedShift:

RedShift is a cloud based, petabyte-scaled data warehouse service that is available with Amazon AWS. Unlike a relational database, it holds a column oriented DBMS that is capable of handling large scale data-sets with parallel processing. Using the Attunity Replicate and Attunity CloudBeam combines well together which accounts for the flexibility to change the configurations based on the data needs and thereby achieve high speed performance. Since the Redshift runs on Amazon AWS, it provides additional advantage of no hardware/maintenance cost.

HauteLook, a Nordstorm acquired company, uses Amazon RedShift for its cloud BI solution. It is reported that the company was able to improve its speed of processing and at the same time maintain the cost of the data warehousing by moving to Amazon Redshift.

2.       P&G with Teradata Enterprise Analytics on Cloud:

Teradata offers Enterprise Analytics on Cloud. It offers different variations in the cloud services as seen in the architecture: Private Cloud, Public Cloud and a Hybrid.

Teradata provides a Unified Data warehouse architecture using Discovery as a service and Data Management as a service on cloud. The Discovery platform uses MapReduce functions drives the sophisticated path and graph analysis on multi structured data to enable rapid exploration and discovery of data. Data Management module capitalizes Hadoop for large-scale dataset management which enables better analysis and management of data of any size.

P&G, a world-class leader in consumer product industry, uses Teradata Enterprise Analytics on Cloud for their global marketing. P&G has incorporated Teradata Customer Interaction Manager, Teradata Master Data Management and Teradata Warehouse miner. Using these the company was able to drill down and analyze their global marketing campaigns to find out their effectiveness. The company also talks about the speed of deployment and low capital costs incurred to set up this analytics tool. The company achieved creating conversations in cloud with about 4.8 billion consumers and was able to maintain all this data in one fresh location available for use by all P&G around the globe.

3.       Coach with MicroStrategy Cloud Business Intelligence Platform:

Microstrategy provides a powerful cloud analytical BI platform provides self-service analytics, big data analytics and mobile analytics. The main advantage is that it allows for combining of information from different systems without the need for any scripting language. It provides powerful predictive, data integration capabilities, mobile support and advanced analytical tools to help customers get the best results. The most differentiating factor for microstrategy is that unlike others, they host their software in their own data center and also allow customers to keep their data on their own premise to alleviate security concerns.

Coach is a leading designer accessories company. Using Microstrategy BI, the company achieved quick BI insights and decision making capabilities. Some of the main features appreciated by Coach are the re-usable report templates, timely visibility to key metrics on dashboard, visibility to granular level of details on the reports and interactive mobile applications.

Resources:

http://aws.amazon.com/redshift/

http://www.microstrategy.com/Strategy/media/downloads/about-us/case-studies/microstrategy-customer-success-story_coach.pdf

http://www.tomsitpro.com/articles/best-cloud-bi-solutions,2-558-5.html

http://www.cio.com/article/2860383/data-warehousing/7-amazon-redshift-success-stories.html

http://www.attunity.com/learning/articles/amazon-redshift-changing-way-companies-utilize-data-warehousing

http://www.teradata.com/News-Releases/2013/Netflix-Selects-Teradata-Cloud-for-High-Performance-Analytics/?LangType=1033&LangSelect=true

http://www.teradata.com/cloud/?LangType=1033&LangSelect=true

http://blogs.teradata.com/scott-gnau/taking-analytics-into-the-cloud/

https://marketplace.cisco.com/catalog/products/3801

http://butleranalytics.com/microstrategy-cloud-bi-review/

Data Visualizations

Data warehouses provides us a logical view of the data from relational databases in a more generalized, consolidated, and multidimensional view.  They model focuses more on providing easier access to data and arriving at better insights about the business to make improved decisions. There is no specific way of modeling a data warehouse for a given business case. There will/could be many different models of data warehouse for same business but with different perspective. The best way is identify the one that suits the client requirement/needs the best. To explain in a better way let us take a closer look at different modeling for different business vignettes. We will be looking at the following businesses: Retail, Telecommunications and Airline Transportation.

          1.       Healthcare:

Fig1: Main Dashboard with customizable metrics shown

Healthcare is a large industry with several loosely coupled operations with numerous physicians, patients, specialists and the management across time and location. The management keeps track of all the hospital level hierarchical activities from equipment supply and maintenance to the performance and tracing across various departments. One of the best ways to visualize the data and metrics is shown in Fig1. The dashboard shows more comprehensive information on the census information for the hospital, patient cycle times, currently waiting and occupied physicians, overall staffing and even details on the usage/availability of patient rooms. Each of the data is shown with its average turnaround metric which provides a quick idea of the stance of the hospital. This dashboard could also allow you to add extra metrics/data thereby providing customizable dashboard options. The above data is concerning hospital management. 

Fig2: Drill-down Dashboard for each department workflow level

Drill down options should be provided with the dashboard in order to efficiently manage the workflow in department. As a sample, we can see the patient workflow in a specific department in the image in Fig 2. Complete visualization is through pie chart, line chart, bar graphs. Information like cycle time for a patient visit in this department, time spend over different procedures, service distribution over time are some of the important graphs shown which provides immense depth of details on performance of a department/physicians.

           2.       Telecommunication:

Fig3: Dashboard for Telecom industry operations

Telecommunication business encompasses a number of business operations like installing the hardware, setting up the network, service activation and upgrades, generating bills, receive payments and repair work orders. From the business perspective, the some important insights required could customer churn rate, customer addition by region by date, successful customer support, successful service plans and upgrades, total installation costs by region by date etc.

The best way to visualize telecommunication data is have divide the processes separately and design a graphical view describing each operation. A sample reference visualization is shown in the figure below. The financial indicators are grouped together and shown as line charts like: Total customers by date, financial growth over time, etc. Line charts could also be used to see the performance of the service plans offered and overall performance with competitors. Bar graphs providing information on customer satisfaction with respect to each service provided and plan upgrades happening over time.

Drill down options could be provided here too to see region wise new customer growth, region wise new service contracts signed and service plan distribution details. A consolidated table showing the current state of the company with revenue, average return per user, overall growth rate could all give better insights into the data.

     3.       Transportation- Airline:

The Airline business keeps track of several data like the flight details, passenger information, payment processing, miles traveled, cost incurred for maintenance, flight construction/purchase, fueling costs etc. One of the striking ways represent data in this type of business is through bar charts and line charts. Performance lag could be shown by making two graphs for every metric, target and actual. Individually divided sections clearly showing overall performance/growth, fuel and other costs incurred, customer data, revenue information segment and flight information section will be very useful. Customer section showing line charts for customer satisfaction level, customer booking medium, customer base growth and average mileage points claimed statistics give detailed insights on the market reach and stand of the company.

Fig4: Dashboard for Flight operations

      While the financial section shows revenue, profit margin and market stand among competitors line charts, the flight section could show information about delays by airport, flight cancels by airport, miles flown with passenger load bar charts. There could be an overall performance section which shows information related to top satisfied flights, routes having the most issues, rating the reasons for disruption of flight and destination with maximum profits as bar charts/pie charts. These would provide the company an overall view to make predictions and business decisions like increased flight number on a particular route, improve customer support, change contract with airports, etc. which will have great positive effects on business revenue.

Resources:

http://www.versustech.com/images/Dashboard-ED-lg.jpg

http://www.jship.org/wp-content/uploads/2012/09/Figure-2-Ophthamology-Patient-Flow.png

http://healthdataintel.com/wp-content/uploads/CFO-2.jpg

https://www.cbsi-corp.com/wp-content/uploads/2012/01/BI-Sales-Dashboard-JE.png

http://www.bisoftwareinsight.com/wp-content/uploads/2014/01/Arcplan-1.jpg

https://www.inetsoft.com/images/screenshots/business_dashboard_with_mapping.png

http://www.dundas.com/

http://dashboardspy.com/